As evident from the publication of NIH Guide PA-02-112 "Genetic Epidemiology of Substance Use Disorders," substance use disorders (SUDs) pose serious public health and societal problems. Gene-gene and gene-environment interactions are potentially important factors in understanding the genetic epidemiology of SUDs. Advanced data analysis and modeling techniques become indispensable in this endeavor. Analytic advancement is generally regarded as one of the greatest challenges in genetic dissection of complex disease. To address this need, NIH Guide PA-02-112 encourages development of statistical tools and analytic methods to enhance our understanding of complex phenotypes including SUDs. This application response to this specific programmatic need by developing statistical methods that is useful for unraveling the genetic basis of complex disorders. Specifically, our primary aim is to develop, evaluate, and apply new statistical models (e.g., latent variable models and tree-based models), methods, and software to conduct genetic analyses of complex traits. We are particularly interested in ordinal traits because methods and software virtually do not exist. Once the methodologies are established, companion software will be developed for all of these models and made available to the public on Dr. Zhang's website. While the methodologies are being developed, as Dr. Zhang's group has demonstrated in the past, we will apply them to real data to address important public health problems, for example, those pertinent to the objectives of NIH Guide PA-02-112. We include three databases, which will allow us to study a variety of issues. For example, we will examine the potential sex difference in familial transmission of drug use and alcoholism, and identify candidate genes, gene-gene and gene-environment interactions for nicotine dependence. In our analyses, we will consider multiple phenotypes including alcohol, drug, and tobacco use as well as comorbid psychiatric conditions such as anxiety. Although the focus of application is on genetic epidemiology of SUBs, our methodologies will be useful for understanding complex ordinal traits in general.